This project demonstrates a leader election algorithm using a ConfigMap in Kubernetes. The algorithm ensures that only one pod becomes the leader and performs specific tasks while other pods wait for their turn.
- Kubernetes cluster
kubectlcommand-line tool- Python 3.6 or later
1. Clone the repository:
git clone <repository_url>2. Navigate to the project directory:
cd leader-election-k8s-configmap3. Install the required Python packages:
pip install -r requirements.txtMake sure you have a running Kubernetes cluster and the kubectl command-line tool properly configured to access the cluster.
Deploy the leader election application:
kubectl apply -f deployment.yamlMonitor the logs of the pod to observe the leader election process:
kubectl logs -f <pod_name>Replace <pod_name> with the name of the pod running the leader election application.
Modify the ConfigMap named leader-election to trigger a change in leadership:
kubectl edit configmap leader-electionThis will open the ConfigMap in your default text editor. Modify the leader field to change the leader. Save and close the file.
Note: Ensure that the ConfigMap leader-election exists in the same namespace as the application.
Observe the logs of the leader election application to see the leader change and pod restarts.
To clean up the deployed resources, run the following command:
kubectl delete -f deployment.yamlThis will delete the Deployment, and Kubernetes will automatically remove the associated pods.
If the leader election process or pod restarts are not working as expected, you can check the following:
Verify that the MY_POD_NAME and MY_POD_NAMESPACE environment variables are properly set in the pod. Check the logs of the pod for any error messages or issues. Ensure that the ConfigMap leader-election exists in the same namespace as the application. If you encounter any issues or have questions, please feel free to reach out for assistance.
Happy leader election in Kubernetes!